suspicious - Story 145314 | Emotion Vectors in Gemma4-31B

"What was that?" Maria asked, her voice sharper than she intended. She'd been in the laundry room when she heard a flurry of rapid, staccato words coming from the living room. Leo had been on the phone, she knew, but she hadn’t recognized the language. She'd assumed it was another business call. Now, she wasn't so sure.

Leo was in the living room now, fiddling with the remote. "Nothing," he said, his voice clipped. He avoided her gaze. He had been acting off, lately, a little distant, a little preoccupied. He seemed to be spending more time on the phone, whispering.

She advanced into the room, studying his posture. The way he was hunched, shoulders tense. The way he'd instantly turned away from her. His demeanor did not alleviate her anxieties. He was hiding something. The air in the room felt suddenly thick, hard to breathe.

Emotion: suspicious

Cluster: Suspicion / Vigilance

PC1 (Valence): -0.82 Negative

PC2 (Disposition): 0.82

Role in Research

This story is one of 1,000 stories generated for the emotion suspicious. During extraction, it was fed through Gemma4-31B and its hidden state activations were captured at 11 layers.

The mean activation across all 1,000 suspicious stories, after denoising with neutral dialogue baselines, produces the suspicious emotion vector -- a direction in the model's 5,376-dimensional representation space.

Logit Lens (Layer 40)

Tokens promoted/suppressed when the suspicious vector is projected through the unembedding matrix.

Promoted:

`🤔`	0.420
`🤔`	0.415
`🧐`	0.362
`疑`	0.352
`suspicious`	0.352

Suppressed:

`and`	-0.307
`alleviate`	-0.301
`ได้รับการ`	-0.298
`allevi`	-0.285
`mouthful`	-0.284